The forbidden conclusion: when the data shows one thing, but the conclusion is dictated from above. http://jamanetwork.com/journals/jama/article-abstract/378571 …pic.twitter.com/bo32uACWzo
-
Predictive validity seems off; evidence seems to indicate that medical outcomes from woman doctors are equal or better
It's common to find women doing better than test scores predict. They have some other personality advantages. Shows up like predictive bias.
-
-
Same is seen for SAT for predicting college GPA, as I recall. Here's a simulation. http://emilkirkegaard.dk/understanding_statistics/?app=test_bias_omitted_variable_bias …
-
Yeah I agree re SAT, but my recollection (which plausibly totally wrong) is this goes away when adding school performance to the regression
End of conversation
New conversation -
-
-
That should be controlled for (partially) by school performance
-
Yeah, goes away when you add high school GPA (again, IIRC). I advocate using any valid predictor, not just IQ-like tests.
-
I understand that view; fact that it doesn't go away in the med regression with controls suggests other problems
-
Would be nice with a bunch of big public datasets people could examine. Including objective measures (e.g. patient death).
-
Multiple regression can be highly misleading here. Plausible that doctors with good reputation would have more patient death, for example
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.